JP2010249685A - Flaw detector, flaw detection method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a flaw detector for detecting micro flaws generated on the surface of a metal plate and enabling an inspector to easily recognize the degree of harmfulness thereof, a flaw detection method and a program. <P>SOLUTION: A flaw detector 10 detects micro flaws generated on the surface of a metal plate F which are difficult to be detected by visual observation. The flaw detector 10 includes a flaw candidate extraction part 110 for extracting flaw candidates which are candidates of micro flaws based on at least one of a regular reflection image and an irregular reflection image obtained by photographing the surface of the transported metal plate F; a classification part 120 for classifying a plurality of flaw candidates into a plurality of groups based on a combination of a regular reflection brightness value and an irregular reflection brightness value corresponding to each flaw candidate; and a total number calculation part 140 for calculating the total number of flaw candidates of each of the plurality of groups for each unit area of a predetermined size of the metal plate surface. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a wrinkle detection device, a wrinkle detection method, and a program for detecting minute wrinkles generated on a metal plate surface.

  Various types of metal plates are used in various fields, and the importance of quality control at the time of manufacturing the metal plates is increasing. On the other hand, in the process of manufacturing a metal plate, unexpected flaws may occur on the surface of the metal plate, and it is very important to detect such wrinkles in quality assurance of the metal plate. . In addition, with the technical improvement of products using metal plates and the development of science and technology, the size of wrinkles allowed for metal plates tends to become smaller.

In order to detect wrinkles formed on the surface of the metal plate, a wrinkle detection device is sometimes used. The purpose of this wrinkle detection device is to guarantee the quality of the metal plate, for example, in the case of a steel plate, it is a problem at the customer's site. It was to detect large wrinkles such as beards, slivers, and poor alloying. Such large flaws, for example, or have several cm or more in length and width, or has a number cm ² or more areas, each one of the flaws was detrimental. Therefore, the inspector visually confirms the detection result of the flaw detection device and the actual flaw on the production line or the inspection line, determines whether the shipment is possible, changes the shipping destination from the original plan, Processing such as cutting the buttock and cutting the buttocks.

  It is very important for the inspector to take the above actions for such a large bag, and it is very important to check the shape and size of the bag. Various techniques have been developed for the display method and wrinkle detection algorithm.

  The eyelid detection device disclosed in Patent Document 1 is used when the subject corresponding to the eyelid position reaches a predetermined position in order to facilitate the comparison between the actual eyelash and the detection result in the operation line and the inspection line. And display means for displaying wrinkle information relating to the wrinkle position. And this wrinkle detection device displays the approach of the wrinkle position as a captured image or graphic image of the operation line, and moves the coil development view while tracking the plate through the steel plate, so that the wrinkle approaches Make the inspector visually recognize that you are doing. In addition, the wrinkle detection device disclosed in Patent Document 2 has a method of displaying wrinkles generated periodically on a map (coil development view).

Japanese Patent Laid-Open No. 5-149892 JP 2005-241356 A Japanese Patent Application Laid-Open No. 11-21174

  On the other hand, as described above, the allowable wrinkle size tends to be small, and in recent years, the number of cases where a wrinkle having a size of about 0.5 mmφ is not allowed is increasing. Such wrinkles are referred to herein as “micro wrinkles”. The size of the wrinkles that are regarded as minute wrinkles varies depending on the usage of the metal plate and the like, and also varies depending on the conveyance speed, the time that the inspector can spend on inspection, and the like. However, in reality, it is difficult for an inspector to visually detect such minute wrinkles from a metal plate conveyed on the line.

  Such small wrinkles often do not pose a problem even if they occur independently of large wrinkles (harmless). However, when a large number of minute wrinkles are generated in a dense manner, the allowable range of wrinkles may be exceeded, which may be problematic (harmful). Therefore, conventionally, as shown in FIG. 13, in an operation line, a detection line, or the like, a predetermined detection range Ar (length of about 5 m, etc.) on the surface of the metal plate F is set at a plurality of locations (about 2-3) at predetermined intervals. In many cases, quality assurance is performed by stopping the plate of the metal plate F at each location and visually observing it using a loupe or the like.

  As described above, it is difficult to visually detect the minute wrinkles. However, if the detection is performed with the line stopped as described above in order to find the minute wrinkles, the metal plate manufacturing efficiency decreases. In addition, the burden on the inspector is very large. Moreover, in the detection by sampling as described above, it is difficult to inspect the entire metal plate, and there is a possibility that a minute wrinkle or the like not in the detection range may be missed. However, as described above, since most of the wrinkle detection devices are intended to detect large wrinkles, it is also a fact that it is difficult to detect such minute wrinkles.

  For example, the wrinkle detection device described in Patent Document 1 is designed to automatically stop a line when a single large wrinkle approaches, and efficiently collate large wrinkles, and detect minute wrinkles. Difficult to do. In particular, as described above, the generation density and the number of minute wrinkles are more problematic than one by one. However, it is difficult for the wrinkle detection apparatus described in Patent Document 1 to detect minute wrinkles, and it is difficult to provide guidance to the inspector about the density and the generation area. Therefore, the actual situation is not necessarily suitable for detecting minute wrinkles.

  In addition, the wrinkle detection device described in Patent Document 2 displays a wrinkle generation region that occurs periodically for large wrinkles. Therefore, as with the wrinkle detection device described in Patent Document 1, it is difficult to detect minute wrinkles, and it is also difficult to provide guidance to the inspector on the density and generation area of minute wrinkles. Moreover, since special processing specialized for periodic wrinkles is required, the processing becomes complicated.

  On the other hand, although it is still difficult to detect minute wrinkles, a wrinkle detection device that handles the generation of large wrinkles as a set is disclosed in Patent Document 3 described above.

  The wrinkle detection device described in Patent Document 3 determines the wrinkle type based on the characteristic parameters of the respective wrinkle parts that have been discriminated and extracted, calculates the distance between the wrinkles determined to be the same wrinkle type, Connect the ridges within the distance to a single ridge, calculate the characteristic parameters of the ridge after linking, and determine the heel grade based on this calculation result

  As described above, the wrinkle detection device described in Patent Literature 3 determines the wrinkle type from a single wrinkle image and expresses the feature amount based on the wrinkle type in order to express the feature amount of densely generated wrinkles. Is recalculated. However, when this apparatus is applied to the detection of minute wrinkles, since the minute wrinkles are small, it is difficult to extract significant feature parameters of individual wrinkles from the captured image of the metal plate, and the accuracy of distinguishing the wrinkles is lowered. In addition, since the distance calculation is performed after determining the type of each bag based on the characteristic parameter, the calculation time becomes enormous, and it is still possible to detect and provide guidance by determining in real time using the line passing speed of the line. difficult.

  Accordingly, the present invention has been made in view of the above problems, and an object of the present invention is to detect minute wrinkles generated on the surface of a metal plate and easily recognize the harmfulness thereof to an inspector. There is to make it.

In order to solve the above problems, according to one aspect of the present invention, there is provided a wrinkle detection device that detects minute wrinkles that are difficult to detect visually on a metal plate surface,
Based on at least one of a specular reflection image and a diffuse reflection image obtained by imaging the surface of the metal plate to be conveyed, a wrinkle candidate extraction that extracts a wrinkle candidate that is a candidate for the minute wrinkle included in the image And
Each of the plurality of wrinkle candidates extracted by the wrinkle candidate extraction unit is divided into a plurality of groups based on a combination of a regular reflection luminance value in the regular reflection image and a diffuse reflection luminance value in the irregular reflection image corresponding to the wrinkle candidate. A classification unit for classifying into
A total number calculating unit that calculates the total number of the wrinkle candidates for each of the plurality of groups for each unit region of a predetermined size on the surface of the metal plate;
A wrinkle detecting device is provided.

  Further, the classification unit is based on a combination of high and low comparing the specular reflection luminance value and the irregular reflection luminance value on the surface of the metal plate with respect to a specular reflection maximum threshold and an irregular reflection maximum threshold separately determined, respectively. Each of the plurality of wrinkle candidates may be classified into the plurality of groups.

Further, for each of the plurality of groups, an importance setting unit that sets an importance coefficient that is an index of the importance of minute wrinkles classified into the group,
Based on the total number calculated for each group and each unit area by the total number calculation unit and the importance coefficient of the group corresponding to the total number, the occurrence degree of the wrinkle candidate is calculated for each unit area An occurrence degree calculation unit to perform,
You may have.

  In addition, for each group and each unit area, the occurrence degree calculation unit adds a value obtained by multiplying or dividing the total number of the groups by the importance coefficient of the group to all the groups. The occurrence degree of the wrinkle candidate may be calculated for each unit area.

  Moreover, you may further have a display control part which displays on the display screen for every said unit area, expressing the degree of occurrence which the said degree-of-occurrence calculation part calculated with the color or the shading.

Also, an occurrence degree recording unit in which the occurrence degree calculated by the occurrence degree calculation unit is recorded,
A time-dependent change creation unit for creating a graph representing a change in the occurrence degree with respect to time;
May further be included.

Further, for each of the plurality of groups, further comprising a dense cocoon identification unit that identifies a plurality of cocoon candidates dense from the plurality of cocoon candidates based on distances between the plurality of cocoon candidates included in the group,
The total number calculating unit may calculate the total number of the plurality of wrinkle candidates specified by the dense wrinkle specifying unit for each of the plurality of groups as the total number of the wrinkle candidates for each of the plurality of groups. .

  Further, the wrinkle candidate extraction unit may extract a fine wrinkle having a vertical and horizontal size that is not more than eight times the pixel size of the image as the wrinkle candidate.

In order to solve the above problem, according to another aspect of the present invention, there is a wrinkle detection method for detecting minute wrinkles that are difficult to detect visually on a metal plate surface,
Based on at least one of a specular reflection image and a diffuse reflection image obtained by imaging the surface of the metal plate to be conveyed, a wrinkle candidate extraction that extracts a wrinkle candidate that is a candidate for the minute wrinkle included in the image Steps,
Each of the plurality of wrinkle candidates extracted in the wrinkle candidate extraction step is divided into a plurality of groups based on a combination of a regular reflection luminance value in the regular reflection image and a diffuse reflection luminance value in the irregular reflection image corresponding to the wrinkle candidate. A classification step to classify into,
A total number calculating step of calculating the total number of the wrinkle candidates for each of the plurality of groups for each unit region of a predetermined size on the surface of the metal plate;
There is provided a method for detecting wrinkles.

In order to solve the above problem, according to another aspect of the present invention, there is provided a program for causing a computer to realize a wrinkle detection function for detecting minute wrinkles that are difficult to detect visually on a metal plate surface. There,
On the computer,
Based on at least one of a regular reflection image and a diffuse reflection image obtained by imaging the surface of the metal plate to be conveyed, a wrinkle candidate extractor that extracts a wrinkle candidate that is a candidate for the minute wrinkle included in the image Function and
Each of the plurality of wrinkle candidates extracted by the wrinkle candidate extraction function is divided into a plurality of groups based on a combination of a regular reflection luminance value in the regular reflection image corresponding to the wrinkle candidate and a diffuse reflection luminance value in the irregular reflection image. A classification function to classify
A total number calculation function for calculating the total number of the wrinkle candidates for each of the plurality of groups for each unit region of a predetermined size on the surface of the metal plate;
A program for realizing the above is provided.

  As described above, according to the present invention, the wrinkle candidates can be classified into a plurality of groups based on the specular reflection luminance value and the irregular reflection luminance value of each of the wrinkle candidates, and the wrinkle candidates in each group for each predetermined region The total number can be calculated. Therefore, the inspector can recognize the harmfulness of the wrinkle candidate included in the group based on the total number or the like based on the total number or the like. Therefore, according to the present invention, it is possible to detect minute wrinkles generated on the surface of the metal plate, and to make the inspector easily recognize the degree of harmfulness.

It is explanatory drawing for demonstrating an example of the line by which the wrinkle detection apparatus which concerns on 1st Embodiment of this invention is arrange | positioned. It is explanatory drawing for demonstrating the structure of the wrinkle detection apparatus which concerns on the same embodiment. It is explanatory drawing for demonstrating the classification | category by the classification | category part based on the embodiment. It is explanatory drawing for demonstrating the classification | category by the classification | category part based on the embodiment. It is explanatory drawing for demonstrating crowded crowd identification by the crowded crowd identification part which concerns on the same embodiment. It is explanatory drawing for demonstrating operation | movement of the wrinkle detection apparatus which concerns on the same embodiment. It is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on the same embodiment displays. It is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on the same embodiment displays. It is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on the same embodiment displays. It is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on the same embodiment displays. It is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on the same embodiment displays. And FIG. 14 is an explanatory diagram for explaining a configuration example of a computer that realizes a series of processes by executing a program. It is explanatory drawing for demonstrating the micro wrinkle detection method performed by an inspector's visual observation.

  Exemplary embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

  The wrinkle detection device according to the first embodiment of the present invention detects, for example, a fine wrinkle formed on the surface of a metal plate such as a steel plate manufactured in the steel industry or a plate of other non-ferrous metal material. The degree of harm can be determined. As described above, the size of the minute ridge varies depending on the use application of the metal plate and the like, and also varies depending on the conveyance speed and the time that the inspector can spend for the inspection. And although it is difficult for an inspector to visually detect such a fine wrinkle from a metal plate conveyed on the line, the wrinkle detection device according to the first embodiment of the present invention has such a fine wrinkle. Can also be detected to further determine the degree of harm. The size of the minute wrinkle to be detected by the wrinkle detection device according to the first embodiment of the present invention is not particularly limited as long as it is difficult to detect visually, but for example, the vertical and horizontal sizes are those of the captured image. It is desirable that each pixel has a size of 8 times or less. When the vertical and horizontal sizes each exceed 8 times the pixel size, it is possible to extract individual wrinkle parameters from each wrinkle image. Therefore, in the wrinkle detection device according to the first embodiment of the present invention, It is desirable to treat such wrinkles as large wrinkles, and treat wrinkles of a size smaller than that as fine wrinkles.

  In order to facilitate understanding of the effect and the like of the wrinkle detection device according to the first embodiment of the present invention, here, first, problems and the like at the time of detecting a minute wrinkle will be described, and then a steel plate will be taken as an example. Next, a line where fine wrinkle detection is performed will be described. Thereafter, the wrinkle detection apparatus according to the first embodiment of the present invention will be described in the order of configuration, operation, effect, and the like. In addition, the display example etc. which this wrinkle detection apparatus presents to an inspector will be described together in the description of the operation of the wrinkle detection apparatus. That is, below, it demonstrates in the following order.

<1. Summary of problems when detecting minute wrinkles>
<2. An example of a line where minute wrinkles are detected>
<3. Configuration of Wrinkle Detection Device According to First Embodiment>
<4. Operation of Wrinkle Detection Device According to First Embodiment>
<5. Example of effect by wrinkle detection device according to first embodiment>

<1. Summary of problems when detecting minute wrinkles>
As described above, in recent years, a fine wrinkle having a size of about 2 mm × 2 mm or less, for example, 0.5 mmφ or less, has become an important flaw for quality assurance of metal plates. These minute wrinkles are difficult to visually check each wrinkle during the conveyance of the metal plate, and simply detecting it is more than twice the size of the target minute wrinkles (for example, 0.5 mmφ). It is necessary to use a wrinkle detection device having a high resolution of about 0.25 mm × 0.25 mm or more if it is a wrinkle. And in order to eliminate the detection omission of minute wrinkles, it is desired to inspect the metal plate for the entire length instead of performing a partial inspection.

  In addition, these fine wrinkles tend to occur frequently, and in order to detect and provide guidance in real time according to the conveyance speed of the line, it is necessary to speed up the image processing for detecting the fine wrinkles. In addition, the minute wrinkles often occupy only a few pixels in the captured image, and the pixel size that serves as a standard for extracting parameters such as width and length recognized from the image is as described above. In general, about 8 pixels × 8 pixels (64 pixels) are necessary, but with a parameter such as width and length recognized from a small cocoon image having a small image size, a significant feature cannot be extracted, and the accuracy is high. Since species determination is difficult, it is also desired to perform classification of micro-pox and determination of harmfulness while there are few significant features.

  In addition, micro wrinkles are rarely harmful by themselves, areas and distribution of dense wrinkles in metal plate coils, density, features of dense wrinkles, images of wrinkles and steel plate surfaces, etc. Is desired to be extracted. Furthermore, it is also desired that the inspector recognizes the degree of harmfulness of such minute fistula accurately and easily.

  It is known that when these fine wrinkles occur frequently, the coil of the metal plate that is rejected continues. In this case, it is necessary to improve the operation, such as changing the operation conditions and cleaning the threading roll. To that end, it is possible to detect harmful micro-wrinkles early in the operation line and provide guidance to the inspector, and the time series of micro-wrinkle occurrence status, such as which coils are continuous from the coil in the inspection line. It is also desirable to have the inspector recognize changes.

  The problems at the time of detecting minute wrinkles described here are merely examples, and are not intended to limit the wrinkle detection apparatus according to the first embodiment of the present invention that can solve these problems. I will add that.

<2. An example of a line where minute wrinkles are detected>
First, an example of a line where fine wrinkle detection is performed using the wrinkle detection apparatus according to the first embodiment of the present invention will be described with reference to FIG. FIG. 1 is an explanatory diagram for explaining an example of a line on which the wrinkle detection device according to the first embodiment of the present invention is arranged.

  In the present embodiment, a steel plate is taken as an example of the metal plate F that is a detection target in which minute wrinkle detection is performed, and an operation line and an inspection line of the steel industry are taken as an example as a line through which the steel plate is conveyed. explain. The “operation line” in the steel industry is a line in which the manufactured metal plate F is transported by the transport roll R etc. and scraped to the coil C, and the “inspection line” is the reverse of the coil C scraped by the operation line. This is a line where the inspector U inspects wrinkles and the like generated on the metal plate F.

  The wrinkle detection apparatus 10 according to the present embodiment images the metal plate F transported along the operation line, detects a minute wrinkle from the captured image, and detects the detection result and the harmfulness of the minute wrinkle as an inspector U. The guidance is displayed in an easily recognizable form. At this time, in this embodiment, the wrinkle detection device 10 can display the above guidance not only for the inspector U who performs the inspection on the operation line but also for the inspector U who performs the inspection on the inspection line. is there. However, the partner (inspector or the like) to which the wrinkle detection device 10 provides guidance is not particularly limited.

<3. Configuration of Wrinkle Detection Device According to First Embodiment>
Next, the configuration of the wrinkle detection apparatus according to the first embodiment of the present invention will be described with reference to FIG. FIG. 2 is an explanatory diagram for describing a configuration of the wrinkle detection device according to the present embodiment.

(3-1. Outline of configuration)
As shown in FIG. 2, the eyelid detection device 10 according to the present embodiment includes an illumination unit 11, imaging units 12 </ b> A and 12 </ b> B, a control unit 13, a storage unit 14, display units 15 </ b> A and 15 </ b> B, and a signal processing unit. 100.

  The illumination part 11 irradiates light to the metal plate F to be conveyed. In the present embodiment, the illuminating unit 11 illuminates the metal plate F in a state of being rubbed against the roll R. The position at which the metal plate F is illuminated is not particularly limited. However, when the illumination is performed as in the present embodiment, the roll R supports the metal plate F, so that the roll R flutters or vibrates. Imaging can be performed, and the detection accuracy of minute wrinkles can be improved.

  The imaging units 12A and 12B image the metal plate F illuminated by the illumination unit 11. At this time, the imaging unit 12A is arranged so as to capture light that is specularly reflected (also referred to as regular reflection) from the light emitted from the illumination unit 11. A captured image obtained by imaging the regular reflection light by the imaging unit 12A is also referred to as a “regular reflection image”. On the other hand, the imaging unit 12B is arranged so as to capture light that is diffusely reflected (also referred to as irregular reflection) by the light emitted from the illumination unit 11. A captured image obtained by imaging the irregularly reflected light by the imaging unit 12B is also referred to as a “diffuse reflected image”.

  The illumination unit 11 and the imaging units 12A and 12B are controlled by the control unit 13. The control unit 13 controls illumination by the illumination unit 11 and imaging by the imaging units 12A and 12B at a predetermined timing in accordance with an input from a line control device (not shown) or the like. Note that the imaging units 12A and 12B are controlled so as to capture the same portion of the metal plate F at the same time. And the control part 13 is image pick-up part 12A, 12B etc. so that the fine wrinkle detection may be performed with respect to the whole region of the metal plate F, that is, so that the whole region of the metal plate F may be imaged by the image pick-up parts 12A, 12B. It is desirable to control. The control unit 13 also controls the signal processing unit 100.

  The signal processing unit 100 acquires specular reflection images and irregular reflection images from the imaging units 12A and 12B, extracts minute wrinkles by performing predetermined signal processing on these images, and represents the harmfulness of the minute wrinkles. The information is calculated and recorded in the storage unit 14 (an example of the occurrence degree recording unit). Then, the information recorded in the storage unit 14 is processed into an image and information in a form that can be easily recognized by the inspector U, and displayed on the display screens of the operation line display unit 15A and the inspection line display unit 15B. be able to. At this time, which image / information or the like is to be displayed on which display unit 15A, 15B is controlled by the control unit 13 based on the operation of the inspector U or the input of the line control device or the like.

(3-2. Configuration of signal processing unit)
The signal processing unit 100 will be described in more detail.
As illustrated in FIG. 2, the signal processing unit 100 includes a wrinkle candidate extraction unit 110, a classification unit 120, a dense wrinkle identification unit 130, a total number calculation unit 140, an importance setting unit 150, and an occurrence degree calculation unit. 160, a temporal change creation unit 170, and a display control unit 180.

  The eyelid candidate extraction unit 110 acquires the regular reflection image and the irregular reflection image obtained by imaging the same part of the metal plate F from the imaging units 12A and 12B. Then, wrinkle candidate extraction section 110 extracts a “wrinkle candidate” that is a candidate for a fine wrinkle included in the image from at least one of the images. At this time, the eyelid candidate extraction unit 110 performs predetermined image processing on both the regular reflection image and the irregular reflection image to remove the noise in order to improve accuracy in the eyelid candidate extraction processing and subsequent processing. It is desirable.

  More specifically, first, the wrinkle candidate extraction unit 110 performs, for example, shading correction that corrects luminance unevenness, spatial filtering processing that emphasizes wrinkles whose longitudinal direction is a specific direction, and the like on both images. After this image processing, wrinkle candidate extraction unit 110 performs wrinkle candidate extraction processing on at least one of the images. As this extraction processing, for example, binarization of the image luminance can be mentioned. In the binarization process, the eyelid candidate extraction unit 110 sets both high and low threshold values for the luminance value (for example, an average value) of an area where no wrinkles are formed in advance, and based on this threshold value. To binarize the image. More specifically, the luminance value between the high and low thresholds is set to “0” as “normal”, and the luminance value exceeding the high threshold or less than the low threshold is set to “1” as “abnormal”. As a result, the wrinkle candidate extraction unit 110 can extract the wrinkle candidate. Here, the “cancellation candidate” means a region that takes a luminance value that is expected to be a fine crease formed on the metal plate F, and is not necessarily captured as a fine crease, but a large crease or a stain is captured. What you get will be included.

  Then, cocoon candidate extraction unit 110 performs labeling on each cocoon candidate. However, at this time, the eyelid candidate extraction unit 110 can also remove the eyelid candidates having the number of pixels equal to or less than a predetermined threshold (for example, one pixel) and label the remaining eyelid candidates in order to further reduce noise. It is.

  In addition, for the wrinkle candidate extracted in this way, the wrinkle candidate extraction unit 110 extracts the feature information of the wrinkle candidate from each of the regular reflection image and the irregular reflection image. As the feature information, for example, position information (width direction coordinate and longitudinal direction coordinate) indicating the position of the eyelid candidate (the center of the circumscribed rectangle of the eyelid candidate or the center of gravity of the binary image) on each surface of the metal plate F; And feature amount information representing the feature amount of the wrinkle candidate that can be extracted from each image. Further, as the feature amount information, for example, the width in the width direction coordinate of the heel candidate, the length in the longitudinal coordinate of the heel candidate, the shape in the surface of the heel candidate, the highest brightness, the lowest brightness, and the heel candidate For example, a histogram of luminance distribution. At this time, when the wrinkle candidate extraction process based on one image is performed, the wrinkle candidate in the other image is specified based on the position information extracted from the one image.

  The classification unit 120, for each of the plurality of wrinkle candidates extracted by the wrinkle candidate extraction unit 110, the luminance value (also referred to as “regular reflection luminance value”) corresponding to the wrinkle candidate and the luminance value in the irregular reflection image. (Also referred to as “diffuse reflection luminance value”), the number of classification into a plurality of groups.

The classification process by the classification unit 120 will be described in more detail.
The classification unit 120 performs regular reflection luminance values and irregular reflection on luminance values from positions other than the wrinkle candidates on the surface of the metal plate F (for example, average values may be included, and in this case, the luminance values at the wrinkle candidate positions may also be included). Each of the plurality of wrinkle candidates is classified into a plurality of groups based on the combination of the brightness values. FIG. 3 is an explanatory diagram for explaining classification by the classification unit according to the present embodiment. The wrinkle candidate has irregularities on the other surfaces, and has a different reflectance. Therefore, as shown in FIG. 3, the regular reflection luminance value and the irregular reflection luminance value are lower or higher than the other parts, respectively. FIG. 3 shows an example of the combination of the luminance values. If the reflected light is strong, the wrinkle candidate has a higher luminance value than the surrounding luminance value. Such an increase in luminance value is also referred to as “white” here, and is represented by a white circle (◯) in FIG. On the other hand, when the reflected light is weak, the wrinkle candidate has a weak luminance value compared to the surrounding luminance values. Such a decrease in luminance value is also referred to as “black” here, and is represented by a black circle (●) in FIG. FIG. 3A shows wrinkle candidates that appear white in the regular reflection image and black in the irregular reflection image, and such wrinkles are often poorly plated. FIG. 3B shows wrinkle candidates that are white in the regular reflection image and white in the irregular reflection image, and such wrinkles are often welding dust. In addition, FIG. 3C shows a wrinkle candidate that is white in a regular reflection image and appears in brightness (hereinafter, also referred to as “gray”) in the diffuse reflection image. In many cases, plating is defective. FIG. 3D shows wrinkle candidates that appear black in the regular reflection image and black in the irregular reflection image, and such wrinkles are often attached to dust rolls.

  The cocoon candidate extraction unit 110 extracts feature information of each cocoon candidate. As described above, in the case of a micro cocoon, it is difficult to extract feature information that is significant for specifying the type and degree of harmfulness of the cocoon. Therefore, the classification unit 120 classifies the plurality of wrinkle candidates into a plurality of groups based on such a combination of black and white gray of the regular luminance value.

  FIG. 4 is a table showing combinations of a plurality of groups. FIG. 4 is an explanatory diagram for explaining classification by the classification unit according to the present embodiment.

  When performing this classification, the classification unit 120 uses the regular reflection luminance value and the irregular reflection luminance value as described above. As the regular reflection luminance value and the irregular reflection luminance value, the maximum luminance value (positive maximum luminance and irregular reflection luminance value) is used. Maximum brightness) or minimum brightness values (positive minimum brightness and random minimum brightness) are used. That is, the maximum luminance value is used for the white haze candidate, and the minimum luminance value is used for the black haze candidate. Then, the classification unit 120, for each of the regular reflection luminance value and the irregular reflection luminance value, has a maximum luminance value threshold (positive maximum threshold and irregular maximum threshold) and a minimum luminance value threshold (positive minimum threshold and irregular minimum threshold). ) Is provided separately and compared with this threshold value. As a result, the classification unit 120 classifies the eyelid candidates one by one based on a combination of magnitudes for the respective threshold values of the regular reflection luminance value and the irregular reflection luminance value.

The threshold value of the maximum luminance value (positive maximum threshold value and maximum disturbance threshold value) is determined as follows, for example.
First, the luminance value of the normal part in each of the regular reflection image and the irregular reflection image is determined to be 128 which is the middle of 256 gradations by image processing, and the threshold value of the maximum luminance value is 256 and 128 which are the maximum values of 256 gradations. 192 in between.
Similarly, the threshold value of the minimum luminance value (positive minimum threshold value and irregular minimum threshold value) is determined by determining the luminance value of the normal part in each of the regular reflection image and the irregular reflection image to 128 which is the middle of 256 gradations by image processing. The threshold value of the luminance value is 64, which is an intermediate value between 1 and 128, which is the minimum value of 256 gradations.
In addition, these threshold values and the setting method thereof can be changed depending on the material of the steel plate, the destination of the coil to be shipped, and the like.

  Further, the classification unit 120 sets a threshold value for the area (that is, a threshold value for the number of pixels) for each wrinkle candidate, and performs classification on the area of each wrinkle candidate. In FIG. 4, the area threshold is shown as “white / white threshold, black and white threshold, normal white threshold, black / white threshold, black / black threshold, positive black threshold, white threshold, and black threshold”. Prescribes the maximum area of wrinkle candidates that can be included in each group.

  As a result, the classification unit 120 classifies each of the plurality of wrinkle candidates into groups G1 to G8 depending on whether or not these conditions are satisfied. However, for the wrinkle candidate that does not satisfy any of the conditions, the classification unit 120 determines that it is not a fine wrinkle, excludes those that do not satisfy the luminance value condition, and those that do not satisfy the area condition are large wrinkles. The heel information (captured image, feature information, etc.) of the large heel is recorded in the storage unit 14.

  At this time, it is desirable that the classification unit 120 discriminates the type and degree of the cocoon candidate that is the large cocoon based on the feature information of the large cocoon and records the result in the storage unit 14 as well. At this time, it is possible to determine the type and degree of the large bag by various classification algorithm methods. For example, examples of the determination process include if the rule, neural network, decision tree, support vector machine, and the like. Of course, other classification algorithms may be used. In these classification algorithms, the type and degree of wrinkle candidates can be output with a predetermined algorithm using feature information and the like as input values. In order to establish this classification algorithm, the classification unit 120 receives a plurality of pieces of teacher information including feature information for a wrinkle candidate whose type and degree are known, and a correct determination result for the feature information. Establish an algorithm. Note that the classification algorithm establishment method, the characteristics of each algorithm, and the like are different for each algorithm, and thus detailed description thereof is omitted here.

  For each of the plurality of groups G <b> 1 to G <b> 8, the crowded soot identifying unit 130 identifies a plurality of crowded soot candidates from the plurality of soot candidates based on the distance between the plurality of soot candidates included in each group. The identification of the dense minute folds is performed, for example, by nearest neighbor distance calculation.

  This nearest neighbor distance calculation will be described with reference to FIG. First, one cocoon candidate D0 is selected, and two other cocoon candidates close to the cocoon candidate D0 (two or more may be extracted) are extracted. Then, the distances l1 and l2 from the reference wrinkle candidate D0 to the respective wrinkle candidates D1 and D2 are calculated. Then, the value obtained by dividing the sum of the distances (l1 + l2) by the number of other wrinkle candidates (2) is compared with a threshold L provided in advance. It is determined that the bag is a bag and if the threshold value L is exceeded, it is determined that the bag is not a crowded bag. The crowded soot identification unit 130 identifies such crowded soot by making such a determination for all of the soot candidates. In FIG. 5, dense folds are represented by black circles (●), and fine folds that are not dense folds are represented by white circles (◯). In other words, in the case of non-dense crowding, (l3 + l4) / 2> L, and as a result, it is determined that there is no crowded crowding.

  It should be noted that the processing by the crowded density identification unit 130 is not necessarily performed because a calculation load is applied, and the classification result performed by the classification unit 120 may be directly sent to the total number calculation unit 140. However, as described above, micro cocoons often affect the quality when they are dense. Therefore, when such dense cocoon identification processing is performed, the accuracy of determining the degree of harmfulness of micro cocoons can be improved. Can do.

  The total number calculation unit 140 calculates the total number of wrinkle candidates for each of the plurality of groups G1 to G8 for each unit region of a predetermined size on the surface of the metal plate F. That is, the total number calculation unit 140 first divides the surface of the metal plate F into predetermined unit areas. The unit area can be set as appropriate. For example, the width direction of the metal plate F is divided into 1 to 16 in the width direction, and the longitudinal direction (coil passing direction) is an arbitrary size of 1 to 100 m. It is desirable to be set in the area. However, here, a case will be described in which the metal plate F is set as a unit region in a region divided into eight in the width direction and divided every 10 m in the longitudinal direction. And the total number calculation part 140 calculates the total number N1-N8 of the wrinkle candidate contained in each group G1-G8 for every unit area | region.

  Note that, as described above, the wrinkle candidates counted here may be only minute wrinkles that are determined to be dense wrinkles by the dense wrinkle identification unit 130, and in the group where the dense flaw identification processing is not performed. It may be a fine wrinkle. Further, the dense crowd identification by the dense crowd identification unit 130 may be performed for each divided unit area after the total number calculation unit 140 divides the unit area. Furthermore, the densely crowded identification unit 130 can identify a crowded cluster not only among the candidates for a group within a group but also across groups.

  The importance setting unit 150 sets importance coefficients C1 to C8 that are indices of the importance of the minute wrinkles classified into the groups for each of the plurality of groups G1 to G8. Here, the importance coefficients C1 to C8 are represented by numerical values, and can be used as an index or a weight representing how much the minute wrinkles in the group are regarded as important. The importance coefficients C1 to C8 are preferably set in advance or automatically set appropriately according to the destination of the coil to be shipped.

The degree-of-occurrence calculation unit 160 calculates the unit area based on the total numbers N1 to N8 calculated by the total number calculation unit 140 for each group and each unit area, and the importance coefficients C1 to C8 of the groups corresponding to the total number. The occurrence degree P of the wrinkle candidate is calculated every time. More specifically, for each unit region, the occurrence degree calculation unit 160 multiplies or divides the total numbers N1 to N8 by importance factors C1 to C8 for each group G1 to G8, as shown in the following formula 1. Are added to all the groups G1 to G8 to calculate the occurrence degree of the wrinkle candidate.
P = C1 * N1 + C2 * N2 + C3 * N3 + C4 * N4
+ C5 × N5 + C6 × N6 + C7 × N7 + C8 × N8 (Formula 1)

  Then, the occurrence degree calculation unit 160 records the occurrence degree calculated for each unit area in the storage unit 14. Note that the “occurrence degree” is a numerical value indicating how much minute wrinkles have occurred, and as described above, weighting is added according to the importance of each group, so that the minute groups of important groups It is possible to express 疵 with emphasis. The importance coefficients C1 to C8 may be set, for example, between 0 and 10, and when multiplication is used as in Expression 1, the larger the importance coefficients C1 to C8, the larger the group. The minute wrinkles inside will be emphasized.

  In addition to the occurrence degree P, the occurrence degree calculation unit 160 calculates the total number of wrinkle candidates classified into any group, that is, the number of wrinkle candidates determined to be minute wrinkles, and is 1 m × 1 m. It is also possible to calculate the maximum value (maximum generation density) or the average value (average generation density) for each side of the coil within a predetermined range of generation density. Further, the total number, density, and the like of these are aggregated in units of one side of the coil and recorded in the storage unit 14.

  The temporal change creation unit 170 processes the degree of occurrence P and the total number / density recorded in the storage unit 14 for each coil C of the metal plate F or every predetermined time interval, and the coil of those values is processed. It converts into the data showing every change or time-dependent change (time change), and creates a graph with the data. The graph and the like created by the temporal change creation unit 170 will be described in the operation of the wrinkle detection device 10.

  The display control unit 180 is controlled by the control unit 13 based on the operation of the inspector U and the input of the above-described line control device and the like, or the information recorded in the storage unit 14 or the graph created by the temporal change creation unit 170 Are displayed on the display screens of the display units 15A and 15B. At this time, the display control unit 180 expresses the degree of occurrence of the minute wrinkles especially by the inspector U by expressing the degree of occurrence calculated for each predetermined unit area with colors and shades of the color on the coil development diagram. Display in a form that is easy to recognize. Note that examples of display by the display control unit 180 will be described in the operation of the eyelid detection device 10.

<4. Operation of Wrinkle Detection Device According to First Embodiment>
The configuration and the like of the wrinkle detection device according to the first embodiment of the present invention have been described above.
Next, operations and the like of the wrinkle detection device according to the first embodiment of the present invention will be described with reference to FIGS. FIG. 6 is an explanatory diagram for explaining the operation of the wrinkle detection device according to the present embodiment. 7-10 is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on this embodiment displays.

  When the metal plate F is transported along the operation line, the control unit 13 controls the illumination unit 11 and the imaging units 12A and 12B to image the surface of the metal plate F. And the signal processing part 100 which acquired the regular reflection image and irregular reflection image obtained as a result of imaging process step S01 shown in FIG.

  In step S01, the eyelid candidate extraction unit 110 performs predetermined image processing on the regular reflection image and the irregular reflection image. Further, wrinkle candidate extraction unit 110 extracts a wrinkle candidate including a fine wrinkle based on at least one image luminance, and performs labeling on the wrinkle candidate. Then, cocoon candidate extraction unit 110 extracts feature information for each cocoon candidate. After the process of step S01, the process proceeds to step S03.

  In step S03, the classification unit 120 classifies the plurality of wrinkle candidates into a plurality of groups G1 to G8 based on the combination of the regular reflection luminance value and the irregular reflection luminance value of each wrinkle candidate. Then, the process proceeds to step S05.

  In step S05, the classification unit 120 confirms whether or not the wrinkle candidate is classified into any of the groups G1 to G8 in step S03, and if it is not classified into any of the groups G1 to G8, the minute wrinkle is determined. If not, the process proceeds to step S07. On the other hand, if it is a micro cocoon classified into any of the groups G1 to G8, the process proceeds to step S09.

  If it is determined in step S05 that it is not a fine wrinkle and the process proceeds to step S07, in this step S07, feature information or the like of a wrinkle candidate that is not the fine wrinkle (for example, large wrinkles or dirt) is recorded in the storage unit 14. . At this time, the classification unit 120 preferably determines the type and degree of these wrinkle candidates using a predetermined classification algorithm, and records the determination result in the storage unit 14 as well. After step S07, the process proceeds to step S09.

  In step S09, the eyelid candidate extraction unit 110 confirms whether or not all eyelid candidates have been extracted from at least one of the regular reflection image and the irregular reflection image. If not all the wrinkle candidates have been extracted, the processing from step S01 is repeated, and if all the wrinkle candidates have been extracted, the process proceeds to step S11.

  In step S11, the dense crowd identifying unit 130 identifies the dense crowd from the candidates in the groups G1 to G8 for each group G1 to G8. Then, the process proceeds to step S13. This step S11 may be omitted, or may be performed for each unit area after the process of step S13. Moreover, in step S11, it is also possible to specify crowding over all groups G1-G8 instead of every group G1-G8.

  In step S13, the total number calculation unit 140 divides the metal plate F into a plurality of unit regions. Then, the process proceeds to step S15.

  In step S15, the total number calculation unit 140 calculates the total number N1 to N8 of wrinkle candidates in the groups G1 to G8 for each unit region. Then, the process proceeds to step S17.

  In step S17, the importance setting unit 150 sets the importance coefficients C1 to C8 of the groups G1 to G8, and proceeds to step S19. The occurrence degree calculation unit 160 calculates the occurrence degree P for each unit area. To do. At this time, the generation degree calculation unit 160 calculates the total number of minute wrinkles, for example, the maximum value (maximum generation density) and average value (average generation density) of each generation surface of the generation density in the range of 1 m × 1 m. It is desirable that the total number, density, and the like of these are calculated and totaled in units of one side of the coil. Then, the process proceeds to step S21, and the occurrence degree P, density, and the like are recorded in the storage unit 14. After the process of step S21, the process proceeds to step S23.

  In step S23, the display control unit 180 is controlled by the control unit 13 based on the operation of the inspector U and the input of the line control device or the like, and the generation degree P, for example, a coil having a generation density in the range of 1 m × 1 m. Maximum value (maximum generation density), average value (average generation density) for each side, total number of minute wrinkles on each side of the coil, maximum generation density, average generation density, etc., graph created by the time change creation unit 170, etc. Is displayed on the display screens of the display units 15A and 15B. In addition, it is desirable to display each information in the position which passes in front of the inspector U in synchronism with conveyance of the metal plate F with both display part 15A, 15B.

(4-1. Display example)
Here, display examples that the display control unit 180 displays on the display units 15A and 15B will be described with reference to FIGS. However, the display examples described here are merely examples, and it is needless to say that various modification examples can be considered. 7-11 is explanatory drawing for demonstrating the example of a display which the wrinkle detection apparatus which concerns on this embodiment displays.

  The display control unit 180 according to the present embodiment, for example, a part of the metal plate F that passes in front of the inspector U in synchronization with the transport of the metal plate F on the operation line or the transport of the metal plate F on the inspection line. The detection result of the wrinkle candidate corresponding to is displayed as a display screen as shown in FIG.

  On the display screen (wrinkle detection screen) shown in FIG. 7, “production number” is displayed as the identification number of the metal plate F being conveyed, and the quality and material of the metal plate F are also displayed. The shipping destination of the metal plate F is also displayed as “destination”.

  On the other hand, as described above, the wrinkle detection device 10 detects not only a fine wrinkle but also a large wrinkle, and the type and degree of the large wrinkle, its characteristic information, and the like are stored in the storage unit 14. Therefore, the display control unit 180 displays a development view (coil development view) of the metal plate F on the display screen, and shows the type of large ridges (A, B, etc.) generated at that position on the development view. Display. At the same time, the display control unit 180 displays the regular reflection image and the irregular reflection image of the large bag on the display screen. The bag detection screen of FIG. 7 shows a case where a large bag is updated at a predetermined timing, such as by selection of the inspector U, whenever a bag is found, or when the bag passes in front of the inspector U, etc. In this example, an image of two large bags (the latest bag and the previous bag) is displayed. However, the display method of the large bag is not limited to this example, and for example, the large bag may be displayed as a list.

  As described above, even if one large flaw may affect the quality or the like, the flaw detection device 10 according to the present embodiment recognizes each large flaw when it finds such a large flaw. Guidance is given to the inspector U including predetermined feature information such as the position so that it can be easily performed. As a result, it is possible to make the inspector U recognize the large-sized ridge appropriately, and the quality control of the metal plate F can be performed more effectively.

  On the other hand, as described above, micro wrinkles often affect quality when a large number of micro wrinkles occur, rather than when each occurrence affects quality. Accordingly, as shown in FIG. 7, the display control unit 180 does not provide guidance for each minute wrinkle, but corresponds to each longitudinal position on the right side of the development view of the metal plate F (coil development view). The total in the width direction of the generation degree P of the unit area to be displayed is displayed for each position in the longitudinal direction (density display column). As a result, the inspector U can recognize the generation degree P of the minute wrinkles while recognizing each large wrinkle with reference to the display screen shown in FIG. In FIG. 7, the total number of occurrence degrees P is not displayed as a numerical value, but is displayed as a color or shading corresponding to the numerical value to warn the inspector U. It should be noted that as the degree of occurrence P increases, a display method such as making the color closer to “red” and making the red “darker” is desirable.

  Further, when the “density” button in FIG. 7 is selected by an input device such as a mouse, the display control unit 180 displays the minute wrinkle screen shown in FIG.

  In the minute wrinkle screen shown in FIG. 8, the degree P of occurrence of each minute wrinkle with respect to each longitudinal position displayed on the right side of the development view (coil development view) of the metal plate F in FIG. It is displayed for each unit area or for each unit area. Therefore, the wrinkle detection device 10 can cause the inspector U to easily recognize the location of the metal plate F in which the minute wrinkles are generated by displaying such a fine wrinkle screen. It is. In FIG. 8, it is desirable to display the degree of occurrence P as a color or shade as in FIG. Further, in FIG. 8, the unit area is displayed with the longitudinal resolution similar to that of the coil development view of FIG. 7, but by increasing the resolution, the degree of occurrence of fine wrinkles P and the like are shown in more detail. It is also possible to make the inspector U recognize. In the present embodiment, the case where the wrinkle detection screen of FIG. 7 and the fine wrinkle screen of FIG. 8 are sequentially displayed has been described. For example, a plurality of display screens may be provided to display both screens simultaneously. Is possible.

  Further, when the unit area shown in FIG. 8 is selected by an input device such as a mouse, the display control unit 180 displays a captured image of the unit area as shown in FIG. Therefore, the wrinkle detection device 10 can cause the inspector U to recognize the occurrence state of the minute wrinkles with an actual image, and can appropriately assist the inspector U in quality judgment and subsequent treatment judgment. is there. In addition, although the case where only the regular reflection image is displayed is illustrated in the captured image illustrated in FIG. 9, for example, only the irregular reflection image may be displayed, or the regular reflection image and the irregular reflection image may be displayed simultaneously. Needless to say. In addition, the display control unit 180 determines the degree of harmfulness (grade) and minute wrinkles in the state of occurrence of minute wrinkles according to the occurrence degree P, the group having a large number of wrinkle candidates, the actual number of minute wrinkles, the actual density of minute wrinkles, and the like. It is desirable to display information useful for quality judgment by the inspector U and subsequent treatment judgment, such as the group name (疵 classification), the maximum occurrence density, and the average occurrence density. In addition, identification information (No (Number)) indicating which minute wrinkles or detection area (an example of identification information of the in-coil area: the unit area shown in FIG. 9 is divided into 8 widths at a position of 1110 m in the longitudinal direction from the top of the coil. It means that it is the eighth in the middle.) Etc.

  Therefore, the inspector U who refers to the minute wrinkle screen determines whether or not to perform a detailed visual inspection on the area (unit region) where the minute wrinkles occur, and determines the quality of the metal plate F and the subsequent treatment. Judgment can be made. Therefore, according to the wrinkle detection apparatus 10 according to the present embodiment, the inspector U can not only recognize the occurrence of minute wrinkles, but also can easily recognize the degree of harm.

  On the other hand, an example in the case where the display control unit 180 displays the graph created by the temporal change creation unit 170 on the display screen is shown in FIGS. As described above, the temporal change creation unit 170 converts the degree of occurrence P, the total number, the density, and the like of minute wrinkles into data representing changes or changes with time for each coil, and creates a graph based on the data. Then, the display control unit 180 can display the graph created by the temporal change creation unit 170.

  FIG. 10 illustrates a case where the temporal change creation unit 170 creates a graph by counting the maximum occurrence density of minute wrinkles on one side of the coil every day. In addition, when the data point of each date is selected by an input device such as a mouse, it is desirable that the display control unit 180 display the wrinkle detection screen shown in FIG. 7 or the fine wrinkle screen shown in FIG. On the other hand, the temporal change creation unit 170 may calculate a value such as a maximum generation density or an average generation density for the generation state of minute wrinkles for each coil C, and create a graph based on the values. An example of this graph is shown in FIG.

  It should be noted that the information graphed by the temporal change creation unit 170 (maximum occurrence density, average occurrence density, etc.) is the actual number of wrinkle candidates that the classification unit 120 has determined to be a minute wrinkle in the occurrence degree calculation unit 160 or the like. However, it may be calculated from the degree of occurrence P in which importance factors C1 to C8 are added.

<5. Example of effect by wrinkle detection device according to first embodiment>
The wrinkle detection device 10 according to the first embodiment of the present invention has been described above.
According to this wrinkle detection device 10, it is possible to appropriately classify micro wrinkles, calculate the occurrence degree P based on the classification result, and notify the inspector U. Therefore, it is possible to guarantee quality over the entire length of the fine metal plate F, and it is possible to improve operational productivity and yield by taking early measures such as changing the operating conditions and cleaning the threaded rolls. is there. Furthermore, when calculating the degree of occurrence P, it is possible to adjust the importance coefficients C1 to C8 of the groups G1 to G8 as appropriate, and it is possible to set the species to be detected with priority.

  In addition, a wrinkle candidate including a fine wrinkle is extracted from at least one of the regular reflection image and the irregular reflection image, and the wrinkle candidate is classified only by a combination of the maximum or minimum value of the regular reflection luminance value and the irregular reflection luminance value of the wrinkle candidate. be able to. Therefore, it is possible to classify appropriate wrinkle candidates while reducing the calculation load. As a result, this wrinkle detection device 10 can classify micro wrinkles that can occur in large quantities without impairing real-time performance.

  Further, when displaying the detection result of minute wrinkles, the generation area of the minute wrinkles, the degree of occurrence P, the generation density, and the like can be displayed on the coil development view. Can be notified in a form that is easy to recognize. Furthermore, it is possible to display changes over time in the occurrence state of the minute flaws and changes for each coil, and the inspector U can use this information for investigation of the cause of the occurrence of the minute flaws and planning of countermeasures. It is.

  As mentioned above, although preferred embodiment of this invention was described in detail, referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this example. It is obvious that a person having ordinary knowledge in the technical field to which the present invention pertains can come up with various changes or modifications within the scope of the technical idea described in the claims. Of course, it is understood that these also belong to the technical scope of the present invention.

  The series of processes described in the above embodiments may be executed by dedicated hardware, but may be executed by software. When a series of processing is performed by software, the above-described series of processing can be realized by causing a general-purpose or dedicated computer as shown in FIG. 12 to execute the program.

  FIG. 12 is an explanatory diagram for describing a configuration example of a computer that realizes a series of processes by executing a program. The execution of a program for performing a series of processes by a computer will be described as follows.

  As illustrated in FIG. 12, the computer includes, for example, a bus 901, a CPU (Central Processing Unit) 902, a recording device, an input / output interface 906, a communication device 907, an input device 909, a drive 910, and an output Equipment and so on. These components are connected to each other through a bus 901, an input / output interface 906, and the like so as to be able to transmit information to each other.

  The program can be recorded in, for example, an HDD (Hard Disk Drive) 903, ROM (Read Only Memory) 904, RAM (Random Access Memory) 905, which is an example of a recording device.

  The program is temporarily stored in a removable storage medium 911 such as a magnetic disk such as a flexible disk, an optical disk such as various CDs (Compact Discs), MO (Magneto Optical) disks, and DVDs (Digital Versatile Discs). Can also be recorded manually or permanently. Such a removable storage medium 911 can also be provided as so-called package software. In this case, the program recorded in these removable storage media 911 may be read by the drive 910 and recorded in the recording device via the input / output interface 906, the bus 901, and the like.

  Furthermore, the program can be recorded in, for example, a download site, another computer, another recording device, etc. (not shown). In this case, the program is transferred via a network 908 such as a LAN (Local Area Network) or the Internet, and the communication device 907 receives the program. The program received by the communication device 907 may be recorded in the recording device via the input / output interface 906, the bus 901, and the like.

  Then, the CPU 902 executes various processes according to the program recorded in the recording device, thereby realizing the series of processes. At this time, for example, the CPU 902 may directly read and execute the program from the above-described recording apparatus, or may be executed after being once loaded into the RAM 905. Further, for example, when the program is received via the communication device 907 or the drive 910, the CPU 902 may directly execute the received program without recording it in the recording device.

  Furthermore, the CPU 902 may perform various processes based on signals and information input from the input device 909 such as an imaging device, a mouse, a keyboard, and a microphone (not shown) as necessary.

  The CPU 902 may output the result of executing the above-described series of processing from, for example, a display device 912 such as a monitor, an audio output device 913 such as a speaker or headphones, or an output device such as a lighting device. If necessary, the processing result may be transmitted from the communication device 907 or may be recorded in the recording device or the removable storage medium 911.

  In this specification, the steps described in the flowcharts are executed in parallel or individually even if they are not necessarily processed in time series, as well as processes performed in time series in the described order. Including processing to be performed. Further, it goes without saying that the order can be appropriately changed even in the steps processed in time series.

DESCRIPTION OF SYMBOLS 10 疵 detection apparatus 11 Illumination part 12A, 12B Image pick-up part 13 Control part 14 Storage part 15A, 15B Display part 100 Signal processing part 110 １１０ Candidate extraction part 120 Classification part 130 Congestion determination part 140 Total number calculation part 150 Importance setting part 160 Occurrence Degree Calculation Unit 170 Time Change Creation Unit 180 Display Control Unit Ar Detection Range C Coil F Metal Plate R Roll U Inspector

Claims (10)

A wrinkle detection device that detects minute wrinkles that are difficult to detect visually on the surface of a metal plate,
Based on at least one of a regular reflection image and a diffuse reflection image obtained by imaging the surface of the metal plate to be conveyed, a wrinkle candidate extraction that extracts a wrinkle candidate that is a candidate for the minute wrinkle included in the image And
Each of the plurality of wrinkle candidates extracted by the wrinkle candidate extraction unit is divided into a plurality of groups based on a combination of a regular reflection luminance value in the regular reflection image and a diffuse reflection luminance value in the irregular reflection image corresponding to the wrinkle candidate. A classification unit for classifying into
A total number calculating unit that calculates the total number of the wrinkle candidates for each of the plurality of groups for each unit region of a predetermined size on the surface of the metal plate;
An apparatus for detecting wrinkles, comprising:   The classifying unit is configured to compare the specular reflection luminance value and the irregular reflection luminance value on the surface of the metal plate with respect to a combination of high and low, which are respectively compared with a specular reflection maximum threshold value and an irregular reflection maximum threshold value separately defined. The wrinkle detection device according to claim 1, wherein each of the wrinkle candidates is classified into the plurality of groups. For each of the plurality of groups, an importance setting unit that sets an importance coefficient that is an index of the importance of minute wrinkles classified into the group,
Based on the total number calculated by the total number calculation unit for each group and each unit area and the importance coefficient of the group corresponding to the total number, the occurrence degree of the wrinkle candidate is calculated for each unit area An occurrence degree calculation unit to perform,
The wrinkle detection device according to claim 1, wherein the wrinkle detection device comprises:   The degree of occurrence calculation unit adds, for each group and each unit area, a value obtained by multiplying or dividing the importance number coefficient of the group by the total number of the groups, and adding the value to all the groups. The wrinkle detection apparatus according to claim 3, wherein the degree of occurrence of the wrinkle candidate is calculated for each unit region.   5. The wrinkle detection according to claim 3, further comprising: a display control unit that displays the occurrence degree calculated by the occurrence degree calculation unit in color or shade and displays the unit area on a display screen. apparatus. An occurrence degree recording unit in which the occurrence degree calculated by the occurrence degree calculation unit is recorded;
A time-dependent change creating unit for creating a graph representing a change in the occurrence degree with respect to time;
The wrinkle detection device according to claim 3, further comprising: For each of the plurality of groups, it further includes a dense cocoon identification unit that identifies a plurality of cocoon candidates densely gathered from the plurality of cocoon candidates based on distances between the plurality of cocoon candidates included in the group,
The total number calculating unit calculates a total number of a plurality of wrinkle candidates specified by the dense wrinkle specifying unit for each of the plurality of groups as a total number of the wrinkle candidates for each of the plurality of groups. The wrinkle detection device according to any one of claims 1 to 5.   The cocoon candidate extraction unit extracts a cocoon as a cocoon candidate having a vertical and horizontal size that is 8 times or less of the pixel size of the image, respectively. Detection device. A wrinkle detection method for detecting minute wrinkles that are difficult to detect visually on a metal plate surface,
Based on at least one of a regular reflection image and a diffuse reflection image obtained by imaging the surface of the metal plate to be conveyed, a wrinkle candidate extraction that extracts a wrinkle candidate that is a candidate for the minute wrinkle included in the image Steps,
Each of the plurality of wrinkle candidates extracted in the wrinkle candidate extraction step includes a plurality of groups based on a combination of a regular reflection luminance value in the regular reflection image and a diffuse reflection luminance value in the irregular reflection image corresponding to the wrinkle candidate. A classification step to classify into,
A total number calculating step for calculating the total number of the wrinkle candidates for each of the plurality of groups for each unit region of a predetermined size on the surface of the metal plate;
A wrinkle detection method comprising: A program for causing a computer to realize a wrinkle detection function that detects minute wrinkles that are difficult to detect visually on a metal plate surface,
On the computer,
Based on at least one of a regular reflection image and a diffuse reflection image obtained by imaging the surface of the metal plate to be conveyed, a wrinkle candidate extractor that extracts a wrinkle candidate that is a candidate for the minute wrinkle included in the image Function and
Each of the plurality of wrinkle candidates extracted by the wrinkle candidate extraction function includes a plurality of groups based on a combination of a regular reflection luminance value in the regular reflection image and a diffuse reflection luminance value in the irregular reflection image corresponding to the wrinkle candidate. A classification function to classify
A total number calculating function for calculating the total number of the wrinkle candidates for each of the plurality of groups for each unit region of a predetermined size on the surface of the metal plate;
A program to realize